A billionaire hedge fund manager just admitted his smartest analysts got AI disruption wrong, and his fix is almost insultingly simple.
The Summary
- Dan Loeb's Third Point hedge fund made winning short bets on AI-vulnerable companies, but also badly misjudged which sectors had real moats against disruption, particularly in information services.
- Loeb's solution: mandatory hands-on AI use across the firm, with employees at all technical levels sharing best practices and exchanging ideas on how they're applying it to their work.
- The admission matters because it comes from someone managing billions who thought proprietary data would protect certain businesses. It didn't.
The Signal
Dan Loeb runs Third Point, a New York hedge fund known for aggressive activist positions and deep sector research. When someone at that level says they "thought we knew better" about AI's reach, that's not just humility. That's a market repricing in real time.
Speaking on the "Invest Like The Best" podcast, Loeb said his firm correctly shorted some companies AI was obviously going to eat. But they also held positions in information services businesses they believed had defensible moats. Proprietary datasets. Specialized knowledge. The kind of stuff that's supposed to create pricing power.
"We thought we knew better that AI wasn't really going to affect this part of the infoservices business or these guys had proprietary information."
That assumption cost them. Loeb called it a key "investment lesson" from the past year. Translation: they lost money being wrong about where AI's capabilities actually stop. The interesting part is what he's doing about it.
Third Point is now pushing everyone to use AI tools directly, regardless of technical background. Loeb's team includes computer scientists hired specifically for AI fluency, but he's not creating a specialist team that "does AI" while everyone else watches. He's distributing the learning curve across the entire organization.
Key elements of Loeb's approach:
- Hands-on use for all employees, not just technical staff
- Regular sharing of best practices across teams
- Direct experience as the primary teacher, not training modules or consultants
The logic is almost obvious once you say it out loud. You can't evaluate AI's impact on a business model if you don't viscerally understand what the tools can and can't do. Loeb put it plainly: "The only way to get good at this is just to use it."
The Implication
If a fund manager with access to the best research and smartest analysts got AI adoption curves wrong, what does that say about everyone else's confidence in their own AI-resistant moats? Third Point's mistake wasn't technical. It was experiential. They analyzed AI from the outside instead of learning its texture from the inside.
The takeaway for anyone running a business or managing capital: start using the tools your competitors or disruptors are using. Not in a sandboxed innovation lab. In your actual work. The gap between reading about AI capabilities and feeling them firsthand is where billion-dollar misjudgments hide.